In vitro biocompatibility assessment of multipurpose contact lens solutions: effects on human corneal epithelial viability and barrier function.

PURPOSE: To explore the in vitro effects of multipurpose contact lens solutions (MPSs) on corneal epithelial barrier function and viability. METHODS: Human corneal epithelial cells (HCEpiC) were exposed to 50% MPSs A-G. Viability was determined using metabolic activity, protease release and caspase assays. Barrier function was evaluated using immunostaining for the tight junction protein zonnula occludens-1 (ZO-1) and resistance measurements. RESULTS: MPS A and G did not affect HCEpiC monolayer viability after 2 h, while MPSs B-F significantly decreased viability. There was a significant decrease in stratified HCEpiC viability after exposure to MPSs B-E for 2 h, while there was no effect of MPS A. After exposure of HCEpiC monolayers to MPS A, F or G for 30 min, ZO-1 staining appeared similar to control. HCEpiC exposed to MPSs B and C demonstrated tight junction breakdown. There was no significant change in HCEpiC monolayer resistance after exposure to MPS A or F for 2 h, while MPSs B-E and G reduced resistance. After exposure to MPS A-E, stratified HCEpiC resistance was significantly decreased after 2 or 4 h. The decrease in resistance was significantly less with MPS A as compared to the other MPSs. CONCLUSIONS: MPSs caused varying modifications to cell viability and barrier function in monolayer and stratified HCEpiC. MPS A did not alter monolayer HCEpiC viability or barrier function, while MPSs B-G caused significant decreases of at least one parameter. Furthermore, MPS A had significantly less effect than MPSs B-E on viability and barrier function of stratified HCEpiC.

Effect of a novel multipurpose contact lens solution on human corneal epithelial barrier function.

PURPOSE: To explore the effect of a novel multipurpose contact lens solution (MPS) on the junction protein distribution and barrier function of cultured human corneal epithelial cell monolayers. METHODS: Cultured human corneal epithelial cells (HCEpiC) were exposed to a novel MPS (MPS A; Biotrue™ multi-purpose solution, Bausch & Lomb Incorporated) at 50%, 75% and 100% for 10 or 30 min. Four commercially available MPS products, MPS B (AQuify, Ciba Vision), MPS C (COMPLETE MPS Easy Rub, AMO), MPS D (OPTI-FREE Express, Alcon) and MPS E (OPTI-FREE RepleniSH, Alcon) were tested in parallel. Tight junction structure and integrity were evaluated by confocal microscopy using ZO-1 antibody and scanning microscopy (SEM). Quantitative evaluation of MPSs on epithelial barrier function was determined by measuring transepithelial electrical resistance (TEER) across HCEpiC grown on Transwell Clear permeable supports and on electric cell-substrate impedance sensing (ECIS) electrode arrays. RESULTS: Overall after exposure to the three concentrations (50%, 75%, and 100%) of MPS A, ZO-1 distribution and fluorescent intensity on the cell surface appeared similar to the media control with continuous tight junctions and clear intercellular junctions. At all measured time points after exposure to MPS A (50% or 75%) there was also no effect on the TEER using both resistance methodologies, and SEM showed that MPS A appeared similar to the Hank's balanced salt solution (HBSS) control. In cells exposed to MPS D there was a dose-dependent change in the distribution of ZO-1, some cell detachment, and a decrease in monolayer resistance at all time points measured. Ultrastructurally, MPS D caused gross changes, including damage to cell junctions and plasma membranes. To a lesser extent, the remaining three commercial MPS products demonstrated some effects on tight junction ZO-1 distribution and/or TEER. CONCLUSIONS: Based on the in vitro measurements of tight junction protein expression, monolayer integrity, and transepithelial electrical resistance, the novel multipurpose contact lens solution (MPS A) did not alter corneal barrier function as compared to media, PBS or HBSS control. Clinical significance of the observed differences in epithelial barrier function among the MPSs tested needs further investigation.

Half-lives of plasma membrane Na(+)/H(+) exchangers NHE1-3: plasma membrane NHE2 has a rapid rate of degradation.

The Na(+)/H(+) exchangers NHE2 and NHE3 are involved in epithelial Na(+) and HCO absorption. To increase insights into the functions of NHE2 vs. NHE3, we compared their cellular processing with each other and with the housekeeping isoform NHE1. Using biotinylated exchanger, we determined that the half-life of plasma membrane NHE2 was short (3 h) compared with that of NHE1 (24 h) and NHE3 (14 h) in both PS120 fibroblasts and Caco-2 cells. NHE2 transport and plasma membrane levels were reduced by 3 h of Brefeldin A treatment, whereas NHE1 was unaffected. NHE2 was degraded by the lysosomes but not proteosomes, as demonstrated by increasing levels of endocytosed NHE2 protein after inhibition of the lysosomes, but not with proteosome inhibition. Unlike that of NHE3, basal NHE2 transport activity was not affected by phosphatidylinositol 3-kinase inhibition and did not appear to be localized in the juxtanuclear recycling endosome. Therefore, for NHE2, protein degradation and/or protein synthesis probably play important roles in its basal and regulated states. These results suggest fundamental differences in the cellular processing and trafficking of NHE2 and NHE3. These differences may underlie the specialized roles that these exchangers play in epithelial cells.

Na(+)/H(+) exchanger NHE3 has 11 membrane spanning domains and a cleaved signal peptide: topology analysis using in vitro transcription/translation.

The transmembrane topology of Na(+)/H(+) exchanger NHE3 has been studied using in vitro transcription/translation of two types of fusion vectors designed to test membrane insertion properties of cDNA sequences encoding putative NHE3 membrane spanning domains (msds). These vectors encode N-terminal 101 (HKM0) or 139 (HKM1) amino acids of the H,K-ATPase alpha-subunit, a linker region and a reporter sequence containing five N-linked glycosylation consensus sites in the C-terminal 177 amino acids of the H,K-ATPase beta-subunit. The glycosylation status of the reporter sequence was used as a marker for the analysis of signal anchor and stop transfer properties of each putative msd in both the HKM0 and the HKM1 vectors. The linker region of the vectors was replaced by sequences that contain putative msds of NHE3 individually or in pairs. In vitro transcription/translation was performed using [(35)S]methionine in a reticulocyte lysate system +/- microsomes, and the translation products were identified by autoradiography following separation using SDS-PAGE. We propose a revised NHE3 topology model, which contains a cleaved signal peptide followed by 11 msds, including extracellular orientation of the N-terminus and intracellular orientation of the C-terminus. The presence of a cleavable signal peptide in NHE3 was demonstrated by its cleavage from NHE3 during translational processing of full-length and truncated NHE3 in the presence of microsomes. Of 11 putative msds, six (msds 1, 2, 4, 7, 10, and 11) acted as both signal anchor and stop transfer sequences, while five (msds 3, 5, 6, 8, and 9) had signal anchor activities when tested alone. Of the latter, 3, 5, 6, and 9 were shown to act as stop transfer sequences after C-terminal extension. The actual membrane orientation of each sequential transmembrane segment of NHE3 was deduced from the membrane location of the N- and C-termini of NHE3. The regions between putative msds 8 and 9 and between msds 10 and 11, which correspond to the fourth and fifth extracellular loops, did not act as msds when tested alone. However, the extension of the fifth extracellular loop with adjacent putative msds showed some membrane-associated properties suggesting that the fifth extracellular loop might be acting as a "P-loop"-like structure.

C-terminal domains of Na(+)/H(+) exchanger isoform 3 are involved in the basal and serum-stimulated membrane trafficking of the exchanger.

When expressed either in polarized epithelial cells or in fibroblasts, two Na(+)/H(+) exchanger isoforms, NHE1 and NHE3, have different subcellular distributions. Using a quantitative cell surface biotinylation technique, we found PS120 cells target approximately 90% of mature NHE1 but only 14% of NHE3 to the cell surface, and this pattern occurs irrespective of NHE protein expression levels. In this study, we examined surface fractions of NHE3 C-terminal truncation mutants to identify domains involved in the targeting of NHE3. Removing the C-terminal 76 amino acids doubled surface fractions to 30% of total and doubled the V(max) from 1300 to 2432 microM H(+)/s. Removal of another 66 amino acids increased surface levels to 55% of total with an increase in the V(max) to 5794 microM H(+)/s. Surface fractions did not change with a further 105 amino acid truncation. We postulated that inhibition of the basal recycling of NHE3 could result in the surface accumulation of the NHE3 truncations. Accordingly, we found that, unlike wild-type NHE3, the truncations were shown to internalize poorly and were not affected by PI3 kinase inhibition. However, while the truncations demonstrated reduced basal recycling, they retained the same serum response as full-length NHE3, with a mobilization of approximately 10% of total NHE to the surface. We conclude that basal recycling of NHE3 is controlled by endocytic determinants contained within its C-terminal 142 amino acids and that serum-mediated exocytosis is independently regulated through a different part of the protein.

Short-term regulation of NHE3 by EGF and protein kinase C but not protein kinase A involves vesicle trafficking in epithelial cells and fibroblasts.

NHE3 is an intestinal epithelial isoform Na+/H+ exchanger that is present in the brush border of small intestinal, colonic, and gallbladder Na(+)-absorbing epithelial cells. NHE3 is acutely up- and downregulated in response to some G protein-linked receptors, tyrosine kinase receptors, and protein kinases when studied in intact ileum, when stably expressed in PS120 fibroblasts, and in the few studies reported in the human colon cancer cell line Caco-2. In most cases this is due to changes in Vmax of NHE3, although in response to cAMP and squalamine there are also changes in the K'(H+)i of the exchanger. The mechanism of the Vmax regulation as shown by cell surface biotinylation and confocal microscopy in Caco-2 cells and biotinylation in PS120 cells involves changes in the amount of NHE3 on the plasma membrane. In addition, in some cases there are also changes in turnover number of the exchanger. In some cases, the change in amount of NHE3 in the plasma membrane is associated with a change in the amount of plasma membrane. A combination of biochemical studies and transport/inhibitor studies in intact ileum and Caco-2 cells demonstrated that the increase in brush border Na+/H+ exchange caused by acute exposure to EGF was mediated by PI 3-kinase. PI 3-kinase was also involved in FGF stimulation of NHE3 expressed in fibroblasts. Thus, NHE3 is another example of a transport protein that is acutely regulated in part by changing the amount of the transporter on the plasma membrane by a process that appears to involve vesicle trafficking and also to involve changes in turnover number.

Na(+)/H(+) exchangers (NHE1-3) have similar turnover numbers but different percentages on the cell surface.

NHE1, NHE2, and NHE3 are well-characterized cloned members of the mammalian Na(+)/H(+) exchanger (NHE) gene family. Given the specialized function and regulation of NHE1, NHE2, and NHE3, we compared basal turnover numbers of NHE1, NHE2, and NHE3 measured in the same cell system: PS120 fibroblasts lacking endogenous NHEs. NHE1, NHE2, and NHE3 were epitope tagged with vesicular stomatitis virus glycoprotein (VSVG). The following characteristics were determined on the same passage of cells transfected with NHE1V, NHE2V, or NHE3V: 1) maximal reaction velocity (V(max)) by (22)Na(+) uptake and fluorometery, 2) total amount of NHE protein by quantitative Western analysis with internal standards of VSVG-tagged maltose-binding protein, and 3) cell surface expression by cell surface biotinylation. Cell surface expression (percentage of total NHE) was 88.8 +/- 3.5, 64.6 +/- 3.3, 20.0 +/- 2.6, and 14.0 +/- 1.3 for NHE1V, 85- and 75-kDa NHE2V, and NHE3V, respectively. Despite these divergent cell surface expression levels, turnover numbers for NHE1, NHE2, and NHE3 were similar (80.3 +/- 9.6, 92.1 +/- 8.6, and 99.2 +/- 9.1 s(-1), when V(max) was determined using (22)Na uptake at 22 degrees C and 742 +/- 47, 459 +/- 16, and 609 +/- 39 s(-1) when V(max) was determined using fluorometry at 37 degrees C). These data indicate that, in the same cell system, intrinsic properties that determine turnover number are conserved among NHE1, NHE2, and NHE3.

Fluoroquinolone (ciprofloxacin) secretion by human intestinal epithelial (Caco-2) cells.

1. Human intestinal epithelial Caco-2 cells were used to investigate the mechanistic basis of transepithelial secretion of the fluoroquinolone antibiotic ciprofloxacin. 2. Net secretion and cellular uptake of ciprofloxacin (at 0.1 mM) were not subject to competitive inhibition by sulphate, thiosulphate, oxalate, succinate and para-amino hippurate, probenecid (10 mM), taurocholate (100 microM) or bromosulphophthalein (100 microM). Similarly tetraethylammonium and N-'methylnicotinamide (10 mM) were without effect. 3. Net secretion of ciprofloxacin was inhibited by the organic exchange inhibitor 4,4'-diisothiocyanostilbene-2-2'-disulphonic acid (DIDS, 400 microM). 4. Net secretion of ciprofloxacin was partially inhibited by 100 microM verapamil, whilst net secretion of the P-glycoprotein substrate vinblastine was totally abolished under these conditions. Ciprofloxacin secretion was unaltered after preincubation of cells with two anti-P-glycoprotein antibodies (UIC2 and MRK16), which both significantly reduced secretory vinblastine flux (measured in the same cell batch). Ciprofloxacin (3 mM) failed to inhibit vinblastine net secretin in Caco-2 epithelia, and was not itself secreted by the P-glycoprotein expressing and vinblastine secreting dog kidney cell line, MDCK. 5. Net secretion and cellular uptake of ciprofloxacin (at 0.1 mM) were not subject to alterations of either cytosolic or medium pH, or dependent on the presence of medium Na+, Cl- or K+ in the bathing media. 6. The substrate specificity of the ciprofloxacin secretory transport in Caco-2 epithelia is distinct from both the renal organic anion and cation transport. A role for P-glycoprotein in ciprofloxacin secretion may also be excluded. A novel transport mechanism, sensitive to both DIDS and verapamil mediates secretion of ciprofloxacin by human intestinal Caco-2 epithelia.

Transepithelial transport of the fluoroquinolone ciprofloxacin by human airway epithelial Calu-3 cells.

Although fluoroquinolone antibiotics such as ciprofloxacin are able to gain access to lung tissue and both pleural and bronchial secretions, the characteristics of transport and cellular uptake of ciprofloxacin in human epithelial lung tissue remain obscure. We have chosen human airway epithelial (Calu-3) cells, reconstituted as functional epithelial layers grown on permeable filter supports, as a model with which to assess both transepithelial transport and cellular uptake of ciprofloxacin. Transepithelial ciprofloxacin fluxes in absorptive (apical-to-basal) and secretory (basal-to-apical) directions were similar throughout the concentration range studied (1.0 microM to 3.0 mM). Transepithelial mannitol fluxes measured concurrently were substantially smaller than ciprofloxacin fluxes in Calu-3 epithelia, suggesting the existence of a mediated transcellular route in addition to a paracellular route for transepithelial permeation. Apical-to-basal ciprofloxacin flux (at 10 microM) was inhibited by a 100-fold excess of unlabelled norfloxacin, enoxacin, and ofloxacin, while secretory flux was unaffected. Cellular uptake of ciprofloxacin, determined as a cell/medium ratio, was greater from the basal compartment (2.7-fold) than apical uptake (1.39-fold) measured at 100 microM ciprofloxacin and showed no saturation up to 3 mM ciprofloxacin. Comparison of the permeation of ciprofloxacin was made with that of lipophilic substrates such as vinblastine and digoxin. There was a linear correlation between transepithelial permeability (Pa-b) and their oil/water partition coefficients with mannitol < ciprofloxacin < digoxin < vinblastine. Comparison of transport of ciprofloxacin across human airway Calu-3 epithelia with that across intestinal Caco-2 epithelia emphasizes the absence of a specific secretory pathway; ciprofloxacin permeation in Calu-3 epithelia appears to be mediated primarily by a transcellular route, with mediated transfer at apical and basal membranes occurring via transporters with low affinity to ciprofloxacin.

Transport and epithelial secretion of the cardiac glycoside, digoxin, by human intestinal epithelial (Caco-2) cells.

1. Human intestinal epithelial Caco-2 cells have been used to investigate the transepithelial permeation of the cardiac glycoside, digoxin. 2. Transepithelial basal to apical [3H]-digoxin flux exceeds apical to basal flux, a net secretion of [3H]-digoxin being observed. At 200 microM digoxin, net secretory flux (Jnet) was 10.8 +/- 0.6 nmol cm-2 h-1. Maximal secretory flux (Jmax) of vinblastine was 1.3 +/- 0.1 nmol cm-2 h-1. Cellular uptake of digoxin was different across apical and basal cell boundaries. It was greatest across the basal surface at 1 microM, whereas at 200 microM, apical uptake exceeded basal uptake. 3. Net secretion of [3H]-digoxin was subject to inhibition by digitoxin and bufalin but was not inhibited by ouabain, convallatoxin, and strophanthidin (all 100 microM). Inhibition was due to both a decrease in Jb-a and an increase in Ja-b. Uptake of [3H]-digoxin at the apical surface was increased by digitoxin and bufalin. All cardiac glycosides decreased [3H]-digoxin uptake at the basal cell surface (except for 100 microM digitoxin). 4. The competitive P-glycoprotein inhibitors, verapamil (100 microM), nifedipine (50 microM) and vinblastine (50 microM) all abolished net secretion of [3H]-digoxin due to both a decrease in Jb-a and an increase in Ja-b. Cellular accumulation of [3H]-digoxin was also increased across both the apical and basal cell surfaces. I-Chloro-2,4,-dinitrobenzene (10 microM), a substrate for glutathione-S-transferase and subsequent ATP-dependent glutathione-S-conjugate secretion, failed to inhibit net secretion of [3H]-digoxin. The increase in absorptive permeability Pa-b (= Ja-b/Ca) and cellular [3H]-digoxin uptake upon P-glycoprotein inhibition, showed that the intestinal epithelium was rendered effectively impermeable by ATP-dependent extrusion at the apical surface. 5. A model for [3H]-digoxin secretion by the intestinal epithelium is likely to involve both diffusional uptake and Na(+)-K+ pump-mediated endocytosis, followed by active extrusion at the apical membrane.